Change evidence report
PR 1999 removes a false blanket 30-day claim from the CRM dashboard and labels each headline, secondary value, and sparkline with the period or status its API query actually represents. The evidence compares the same controlled report data at the merge-base and implementation head, then verifies the scope-bearing copy remains visible without horizontal overflow at a 320-pixel viewport.
bb82b0a98 (merge-base with main)…d643958ff (captured implementation head)Reviewer brief
Revenue and bookings identify their last-30-day window, total customers identifies its all-time scope, memberships identifies the API's active-status predicate, every secondary value names its 30-day window, and every sparkline has both a visible Last 7 days caption and a scoped accessible name. The page subtitle now describes mixed-scope metrics without assigning one period to all of them.
Both desktop captures use the same 1440 by 1000 viewport and sanitized controlled API responses. Browser request inspection observed 30-day report queries from 2026-07-27 through 2026-08-25 and seven-day trend queries from 2026-08-19 through 2026-08-25. The narrow capture uses the same proposed response data at 320 pixels. It proves rendering and layout, not production data quality or deployment.
Claim map
Each material changed behavior or state maps to direct evidence or an explicit gap.
| Behavior or claim | Evidence | Status |
|---|---|---|
| Every headline, hint, and trend exposes its true scope | Matched desktop before and after captures, API query inspection, focused render test, and the scope table | Covered |
| The subtitle no longer makes a false blanket 30-day promise | Desktop comparison shows the old blanket statement and the proposed mixed-scope description | Covered |
| Narrow layouts preserve distinguishing scope text | 320-pixel browser capture plus measured 320-pixel document and main scroll widths with no horizontal overflow | Covered |
| Sparkline accessible names state the seven-day period | Browser accessibility inspection and focused assertions for all four data-driven SVG labels | Covered |
Observed change
Evidence is grouped by the reviewer question each item addresses.
CRM dashboard overview
ScreenshotThe merge-base labels only the values and makes one page-level 30-day claim. The proposed dashboard moves the period or status into every headline and hint, and adds a visible seven-day caption above every sparkline.
Responsive overview cards
ScreenshotThe single-column layout preserves long headline and hint text with normal wrapping. Browser measurements reported a 320-pixel viewport, document scroll width, and main scroll width, so the page introduces no horizontal overflow.
API-backed scope contract
TableThe dashboard uses one 30-day summary range, one seven-day trend range, an unbounded customer total, and a status-only membership count.
| Card | Headline scope | Secondary scope | Trend scope |
|---|---|---|---|
| Revenue | Last 30 days | Transactions in the last 30 days | Last 7 days |
| Customers | Total customers, all time | New in the last 30 days | Last 7 days |
| Bookings | Last 30 days | Cancel rate in the last 30 days | Last 7 days |
| Memberships | Records with active status | New and churned in the last 30 days | Last 7 days |
Confidence
15 DashboardPage tests passed with 129 assertions, including the API-backed visible scope contract and all four seven-day accessible summaries
The @sauna-crm/crm typecheck completed successfully
Biome checked the changed DashboardPage source and focused test without diagnostics
The CRM dashboard rendered against controlled report boundaries at 1440 and 320 pixels; every scope was visible, all four SVG names included the last seven days, and narrow document and main scroll widths matched the 320-pixel viewport
Independent review identified an inaccurate now claim for status-only memberships and brittle class-order assertions. Both were corrected before this evidence capture.
Limits
Visible periods follow fixed dashboard presets The dashboard currently requests fixed 30-day summaries and seven-day sparklines. A future preset change must update the visible phrases and focused contract test together.
Captures use controlled report responses Sanitized responses make the before and after layouts directly comparable. API route inspection and focused tests establish semantics separately; the images do not claim production data correctness.